An inverse scheme for sensitivity and uncertainty analysis in basin modeling: the resolution limits of Easy%Ro

Abstract

One of the major contributors to uncertainties in basin modeling arises from the determination of the parameters necessary for reconstructing the thermal history due to the importance of the thermal maturity in evaluating the petroleum potential of a basin. The methods to determine these uncertainties need to be developed, tested and applied. Two major methods (geodynamic models and calibration of paleoheat flow to thermal indicators) are available for determining the paleoheat flux of a sedimentary basin. Of the latter, the chemical kinetic model Easy%Ro (Sweeney and Burnham, 1990) is widely used and has a firm foundation in laboratory experiments and calibration. The Easy%Ro model calculates the systematic variation of vitrinite reflectance with time and temperature. Even though Easy%Ro is widely used for constraining paleoheat flow by calibrating and modeling vitrinite reflectance, its ability to resolve paleoheat flow and its success in providing the relationship between vitrinite reflectance and the thermal history of a sedimentary basin is not yet investigated and determined quantitatively. This study provides the first quantitative approach to determine the resolution limits of the chemical kinetic model, Easy%Ro. Paleothermal gradients are calibrated against vitrinite reflectance using the Easy%Ro model plugged into a single parameter inverse engine in order to perform sensitivity analysis and assess the uncertainty. Vitrinite reflectance data is obtained from the B-1 , Lulu-1 , Mona-1 , Sten-1 and Q-1 wells located in the Danish Central Trough, in the North Sea. A range of geothermal gradients are investigated using the model. As a quantitative measure of mismatch between modeled and measured values, the mean squared residual (MSR) is used (MSR =(1/n) [](Ro[]-Ro[])²). A 90% confidence level on the best answer (lowest MSR) is taken to represent the acceptable error range for the particular model. The sensitivity of the Easy%Ro model to changes in geothermal gradient and its ability to resolve thermal history are investigated from the determined uncertainties associated with scatter in the calibration data (measured vitrinite reflectance). The results are used to elaborate on Easy%Ro resolution limits with respect to thermal history.